Road Network Modeling And Its Path Planning Research Based On Traffic Information

Navigation data models and path planning algorithms are important parts of intelligent transportation system(ITS).At present,there are many data models and path planning algorithms,but there still are some problems in many navigation data models as well as limitations on path planning algorithms.The ant colony optimization(ACO),because of its advantages of strong robustness and easily to be combined with other algorithms,becomes popular among the researchers.However,its slow convergence rate and easily falling into the local optimal limit its wide use.Based on the current status of navigation data model and path planning algorithms,this paper proposed a system of making the path planning.The main achievements and contents are listed below:According to the large number of literatures and based on the current research situation,this paper analyzed the planar navigable data model,lane-based navigable non-planar navigable data model and GIS-T spatio-temporal data model.The basic principle of the ant colony algorithm and its typical development process as well as its application in multi-objective optimization problem were discussed.This paper proposed two methods over the problem of ant colony algorithm.One was to introduce the random factor to accelerate its convergence rate.Random factor drives the ants to make full use of existing information.The other was to introduce the elitist ants and weaken strategy(EAWSA).EAWSA weakened suboptimal path pheromone whenever better path found,alleviating the long-term iterative pheromone accumulation effect.These methods improved the ACO with little algorithm complexity increase.Through the application on TSP problem,the improved algorithm converged faster than the traditional one,along with the better optimal result.In road network processing,SuperMap was used to preprocess the map,including the merging of adjacent endpoints and remove the suspended lines exceeding tolerance.Then,the network search range was limited in the ellipses,whose focuses were the start point and end point specified,reducing the square root calculation.Then we use greedy algorithm to merged fork nodes in branch loops and process the branch loops in the network which cannot be done in the preprocessing,effectively simplified the search range,save the precious pathfinding time.The improved algorithm was transplanted into the actual path planning,including single objective shortest path planning,transform the multi-objective path planning into the single objective path planning and solve the multi-objective path planning based on the Pareto non-inferior solution set.Niche technology was adopted for the purpose of uniformly distributing the solution set on the Pareto frontier.The simulation results showed that the improved algorithm had excellent performance.This paper constructed and simplified the network topology,restricted and simplified the road network,adopted the improved algorithm into reality,and provided a solution of path planning.